Inverse organic-inorganic composite materials. III: High glass content nonshrinking sol-gel composites via poly(silicic acid esters)

Materials Science and Engineering, A162 (1993) 257-264

257

"Inverse" organic-inorganic composite materials: high glass content non-shrinking sol-gel composites Bruce M. Novak and Mark W. Ellsworth Department of Chemistry, Universityof California at Berkeley, Berkeley, CA 94720 (USA)

Abstract We, as well as others, have been interested in the sol-gel process for the synthesis of hybrid inorganic-organic composite materials. Since our first report on the application of tetraalkoxysilanes possessing polymerizable alkoxides for the production of non-shrinking sol-gel composites, we have extended our efforts towards increasing the glass content in these composite materials. The stoichiometry in the tetraalkoxysilanes limits the maximum glass content in the original non-shrinking composites to 10%-18%. In order to increase the glass content to greater than 50%, we focused our efforts on the use of silicic acid oligomers. Molecular weights of the poly(silicic acid) materials were varied from M~ = 5000 to M, = 2 000 000 by controlling reaction conditions. In addition, branching ratios (i.e. linear vs. spherical particles) can be controlled by changing the catalysts used. The properties of the resulting composite can range from a transparent flexible material to a transparent hard material simply by changing the organic polymer in the composite.

1. Introduction

Modern composites embody a general class of materials which is extremely broad, ranging from polymer-polymer blends and reinforced plastics to chopped-fiber and filled polymer composites [1]. The primary properties of structural materials are strength, stiffness and toughness. Secondary considerations include resistance to corrosion, creep, temperature and moisture. Both the strength and stiffness of a composite can be derived from the properties of the reinforcing fiber. Toughness results from the interaction between the matrix and the fibers, thus highlighting the importance of controlling the interfacial properties between the two phases. Occasionally a synergistic relationship exists between the components of a composite. The combination of a brittle fiber in a brittle matrix produces a material which is much tougher than either of the two single components. This synergism is achieved by a combination of mechanisms which tend to keep cracks small and isolated, and which dissipate energy [1]. Composite materials have evolved considerably over time, culminating in today's carbon-fiber reinforced resins [1], carbon-carbon composites [2], aramid fibers [3] (aromatic polyimides) and molecular composites [4]. In the most general sense, maximized mechanical properties should result from the combination of two or more very dissimilar components. A good example of dissimilar compounding is highmodulus glass-fiber reinforced organic matrix materi0921-5093/93/$6.00

als [5]. These glass fiber composites have found a variety of uses ranging from fiberglass for pleasure boats to high-tech, aerospace "stealth" applications. We [6-8], as well as others [9], have been interested in exploring the reverse side of this traditional glassfiber composite approach by using organic polymers to reinforce an inorganic glass matrix. Although "inverted", the basic composite principles remain intact, but now the synergistic relationship results from the combination of a high-modulus organic polymer (for high tensile strength) with a three-dimensionally cross-linked, inorganic matrix (for high compressive strength). In order to achieve this goal, it becomes necessary to form the inorganic matrix under conditions in which organic polymers will survive. This can readily be accomplished using low-temperature sol-gel technology. The sol-gel process for the preparation of glasses under mild conditions has received much attention with respect to the formation of highly homogeneous monolithic glasses and inorganic ceramic composites [10, 11]. The sol-gel process is based on the homogeneous hydrolysis and condensation of metal alkoxides in the presence of cosolvents to form highly cross-linked networks. Under controlled reaction conditions, large-scale, optically transparent monolithic samples can be obtained. The simplest sol-gel process is the formation of SiO2 from the hydrolysis of tetraethoxy orthosilicate (TEOS) (scheme 1 ). A solvent-swollen, three-dimensional SiO2 network is obtained at this point. Controlled drying of the © 1993 - Elsevier Sequoia. All rights reserved

258

B. M. Novak, M. IV. Ellsworth

/

Hydrolysis Si(OR) 4 +

H20

= (RO)3Si-OH + ROH

Condensation \ --Si-OH /

+

/ HO-Si-\

\

/

--Si.O-Si--

/

+

H20

x

and/or N

\ / --Si-O-Si--

/ Si • OH

+

/

RO - Si --

x

/

Net Reaction O"

x

+ ROH

I ~ l - V . s.

,.,. kD I ,

si(OR), n o/I-r or OHL SolventSwollen (liq.) ' Cosolvent - -'si o..u sfV mm SiO2Matrix lb

.S~'O'l

"

.... Si " 0 ~*

Scheme 1

~ ~ .

•

~_ _

H20/H+ )

Si(OR)4

Scheme2 solvent consisting of the excess water, added cosolvent and the alcohol hydrolysis product, under ambient conditions leads to crack-free, monolithic glass formation. One of the major obstacles to the widespread application of sol-gel techniques is the fact that this drying process is accompanied by extraordinary shrinkage of the glass (shrinkages of more than 50% are common)[12]. Sol-gel technology has recently been applied to the formation of composite materials [9]. The most basic approach involves dissolving a preformed organic oligomer or polymer in the sol-gel solution, and then allowing the hydrolysis and condensation of the inorganic network to occur. Under the appropriate conditions, the polymer remains homogeneously embedded in the inorganic gel throughout the synthesis and drying steps (scheme 2).

Inverseorganic-inorganic composites

pregelled, sol-gel solutions. Specifically, poly(2-vinylpyridine), poly(vinylpyrrolidone) and polyacrylonitrile can be dissolved in TEOS-H20 solution (tetramethoxysilane (TMOS) can also be used), using organic acids as cosolvents. Under the proper conditions, subsequent hydrolysis and condensation of the TEOS produces optically clear gels containing the organic polymers. Slow ambient drying results in composite materials in which the organic polymer remains homogeneously embedded within the three-dimensional SiO2 network. These monolithic glassy materials display excellent optical clarity. Without further sintering, however, they remain brittle materials. Further investigations led to the discovery that composites possessing superior mechanical properties could be obtained using cellulosics as the organic component. Using cellulose acetate, optically transparent composites with organic contents ranging from ca. 2% to 85% by weight have been prepared. Scanning electron microscopy (SEM) studies on composites possessing 30% cellulose acetate show a continuous phase of SiO 2 with the organic polymer dispersed in irregular domains averaging ca. over 1/~m in size [13]. Although the mechanical properties of these new composites have yet to be systematically measured, preliminary results indicate that composites with high cellulose contents are exceptionally tough and impact resistant (i. e. vigorous pounding in a mortar and pestle does not shatter these materials). All of the above composites consist of linear polymers dispersed in the inorganic glass matrices without the benefit of planned, covalent links between the two phases. In order to increase further the mechanical properties of these composites, cellulose derivatives were synthesized possessing pendant trialkoxysilane groups (eqn. ( 1 )):

H° o~o

oOd.), n

~Rohs~..~..sco

Results

and discussion

2.1. Preformed polymer composites

During our preliminary studies in this area, we identified a limited number of soluble polymers which, at the conclusion of the condensation and drying processes, remain homogeneously embedded within sol-gel derived SiO2 glasses. For example, we have found that polymers with basic functional groups such as amines and pyridines are soluble in the acid catalyzed,

OH

o

I1

o "~ ~ N . . . . ~ Si(OR)3

.o. II

og+

(1)

Ib

Si(OR)3

2.

HO

0

NH~...~ . 0 - Si-0. ,% Si. • ~ S ' ' 0 Si--

Despite the preliminary success of some of the above materials, the formation of composites by the incorporation of preformed polymers into sol-gel glasses is severely limited by at least two factors. First, only a limited number of polymers are soluble in the tricomponent sol-gel solution. Secondly, the shrinkage associated with their drying introduces a considerable amount of stress within the dried glasses and preludes most molding applications. For these reasons, we began to explore alternative routes into these materials.

B. M. Novak, M. W. Ellsworth / Inverseorganic-inorganic composites

Ultimately, our solutions to the first problem led to solutions to the latter. 2.2. Simultaneous interpenetrating organic-inorganic network composites In order to circumvent the solubility problem associated with trying to incorporate preformed polymers and to provide better homogeneity between the two phases, we began investigations into the formation of simultaneous interpenetrating networks (SIPNs) [14] by the synchronous formation of both the organic polymer and the inorganic glass network [8]. We have identified two organic polymerization methods which are compatible with the restrictive conditions imposed by the sol-gel reaction (i.e. aqueous acidic or basic medium): vinyl, free radical polymerizations and aqueous ring-opening metathesis polymerizations (aqueous ROMP) catalyzed by a variety of Ru 3+ and Ru 2+ salts [15]. A ROMP example of this SIPN process is shown in scheme 3. The utility of this technique is illustrated by the fact that poly-II (scheme 3) is an intractable, totally insoluble polymer and yet, using the SIPN approach, poly-II can be homogeneously embedded within the glass matrix in concentrations up to ca. 60% without undergoing macrophase separation. Verification of polymer formation within these glasses can be accomplished by forming a soluble polymer, crushing the glass after its formation and extracting with a good solvent. For example, poly(5,6-dimethoxymethyl-7-oxanorborn-2ene), poly-III, is formed under these in situ conditions in greater than 95% recovered yield. The molecular weight of poly-llI formed in this in situ process is quite high with Mn = 1.3 x 1 0 6, M w = 2.1 x 106 and PDI = 1.6. In addition to utilizing two independent, non-interfering polymerization techniques, successful SIPN formation requires matching the polymerization rates of the two systems. Significant deviations from these matched rates result in systems which approach the homopolymerization limits: uncontrolled polymer precipitation when the ROMP or free radical rates are greater than the Si(OR)4 condensation rate, or

~,.~

oH

H20

inorganic gells swollen with unreacted monomer when the condensation rate is much greater than the organic polymerization rates. Under ideal reaction conditions, a transparent glass-polymer composite is obtained. SEM studies show that in comparison with the composites formed by incorporating preformed polymers, the SIPN composite materials show far greater homogeneity and small domain sizes (average polymer domain sizes are less than 1000 A). The advantages of the SIPN approach over using preformed polymers to form sol-gel derived composites are three-fold. First, by incorporating difunctional monomers, the SIPN approach allows us to cross-link the organic polymer, thereby locking-in the interpenetrating phase morphology. Secondly, this approach allows for the in situ formation, and hence homogeneous incorporation, of polymers which would normally be completely insoluble. Lastly, these SIPN materials display greater homogeneity and smaller domain sizes than comparable preformed materials. 2.3. Non-shrinking sol-gel composites As with the preformed polymer composites, shrinkage remains a problem in these SIPNs. In an effort to surmount this problem, we have synthesized a series of tetraalkoxysilane derivatives possessing polymerizable alkoxide groups in place of the standard ethoxide or methoxide groups (Table 1 ) [6, 7]. The hydrolysis and condensation of these siloxane derivatives liberates a polymerizable alcohol. In the presence of the appropriate catalyst (free radical or ROMP), and by using a stoichiometric amount of water and the corresponding alcohol as cosolvent, all components of these derivatives are polymerized. Since both the cosolvent and the liberated alcohol polymerize, gel drying is unnecessary and no gel shrinkage occurs. This overall process is illustrated in scheme 4. 2. 4. High glass, non -shrinking composites Since our first report on the application of these tetraalkoxysilanes possessing polymerizable alkoxides for the production of non-shrinking sol-gel composites

~

II or

~

Scheme 3

=

259

SiO 2

260

B. M. Novak, M. W. Ellsworth

/

Inverse organic-inorganic composites

[6, 7], we have extended our efforts towards increasing the glass content in these composite materials. The stoichiometry in the tetraalkoxysilanes limits the maximum glass content in the non-shrinking composites to 10%-18%. In order to increase the glass content to greater than 50%, we focused our efforts toward the use of silicic acid oligomers and/or polymers substituted with polymerizable alkoxides [6]. Poly(silicic acid) can be generated in situ by the hydrolysis and condensation of sodium metasilicate (NazSiO3 •9H2 O) at low pH (3.6 M HCI) and extracted into organic solvent (THF) by the addition of salt to the organic layer [16]. Soluble silicic acid polymers of molecular weights ranging from 8000 to 7 000 000 can be prepared by increasing the reaction time from I h to ca. 72 h (eqn. (2)): .o ? HO

Na2SiO3

1. H20/IICI

2. T

.cl

~Si

,,4.,,

x ~ 60.1 ]JJ 1

\FC/j]

Si , ~ t j ~ OH

OH

0

0~

I~ Si

SI i .

SI i .

SI i .-

Q" -

( o. ( o . " o.)m o . . o . "Q, ,Q, Q3 ?'o.

These preformed polysiloxanes offer two adjustable parameters for controlling glass content: the number of Q3 and Q4 branch points, and the degree of alkoxide substitution. The weight per cent of glass in composites derived from these appropriately substituted polyTABLE 1. Candidatefree radicaland ROMP monomerfor non-shrinkingcomposites ]~OMP Monomers

Free Radical Monomers

0

,,

0

), 0

0

0

0

o o

o" b ~ -

o_o.

~on

[~l

Scheme4

+p

(3)

where MW is the molecular weight of the liberated alcohol, n, m and p are the percentages of Q2, Q3 and Q4 silicon centers respectively, and x and x' are the percentages of alkoxide substitutions on the Qz and Q3 moieties respectively. In addition, control over the Q ratio should allow for the formation of inorganic phases with morphologies ranging from spherical (high percentage of Q 4 ) t o more linear (high percentage of QZand Q3 relative to Q4). We have found that the Q4 content of these silicic acid polymers can be systematically changed from ca. 35% to greater than 70% (as measured by 298i nuclear magnetic resonance (NMR)) by adjusting the HCI concentration between 3 and 6 M. Although highly branched and/or cross-linked, these high Q4 polymers retain their solubility in polar organic solvents. Alkoxide substitution on these poly(silicic acid) polymers can be effected by addition of the appropriate alcohol to the THF solution of polymer followed by an azeotropic distillation to remove the liberated water [16]. By varying the reaction time, the degree of substitution (DS) was controlled by the amount of THF]water azeotrope removed. DS values ranged from 25% to 75% as evaluated by endcapping of the unreacted silanols with trimethylsilylchloride (TMSCI) (scheme 5). Composites were synthesized by allowing the silicic ester to condense in a solution of a polymerizable monomer and a free radical initiator (typically benzoyl peroxide) or aqueous ROMP catalyst. In most cases, transparent inorganic-organic composites are formed. A variety of cosolvent monomers can be used, including cross-linking agents, in order to achieve different properties in the final composite. For example, poly(hydroxyethylacrylate) (HEA) is a rubbery solid at room temperature (T~ = -15 °C) whereas poly(hydroxy-

,1. '~ ,o.

OH

HO;, o